Food

Chemistry
Food Chemistry 104 (2007) 1491–1501
www.elsevier.com/locate/foodchem

The impact of transglutaminase on soy proteins and tofu texture

S. Bin Md Yasir a,b, K.H. Sutton b, M.P. Newberry b, N.R. Andrews a, J.A. Gerrard a,*
a
School of Biological Sciences, University of Canterbury, Christchurch, New Zealand
b
NZ Institute for Crop and Food Research Ltd., Private Bag 4704, Christchurch, New Zealand

Received 9 June 2006; received in revised form 16 December 2006; accepted 23 February 2007

Abstract

The enzyme transglutaminase was investigated for its cross-linking effect on the soy proteins of tofu. In vitro incubations confirmed
that soy proteins are excellent substrates for transglutaminase, especially when denatured. The macroscopic effects resulting from the
addition of transglutaminase were compared to changes at the microstructural and molecular level. Treatment produced a firmer tofu,
with a significantly increased fracture force. Examination by SEM showed a change in the matrix structure, with transglutaminase result-
ing in a finer-stranded, uniform network that accounted for the increase in fracture force. At the molecular level, little, if any, cross-link-
ing occurred within the tofu matrix in situ. This suggests that the change in functional properties afforded by addition of
transglutaminase to tofu is due to a side reaction of the enzyme, for example hydrolysis of glutamine residues, rather than its cross-link-
ing activity. These ideas are further explored in the accompanying paper.
Ó 2007 Elsevier Ltd. All rights reserved.

Keywords: Soy protein; Tofu; Texture; Transglutaminase; Scanning electron microscopy

1. Introduction unit composition of tofu has been shown to correlate

In many foods, proteins play a major role in product
quality and determine many of the functional properties
of these systems (Gerrard, 2002). Our research has previ-
ously demonstrated that cross-linking of wheat proteins
by the enzyme transglutaminase can have a dramatic influ-
ence on the properties of bread (Gerrard et al., 1998) and
croissants (Gerrard et al., 2000), which could be attributed
to cross-linking of specific proteins (Gerrard et al., 2001;
Gerrard & Sutton, 2005). In this research, we sought to
establish whether similar changes of functional properties
in other food products could be correlated with specific
changes in cross-linking patterns.
Tofu was selected as our model because its functional
properties are determined by denaturation of soy proteins
to form a gel. Soy proteins, in their native state, do not
form a gel; they must be heat-denatured and then coagu-
lated to form the tofu (Liu, 1997). Furthermore, the sub-
*
Corresponding author. Tel.: +64 3 3667001; fax: +64 3 3642590.
E-mail address: juliet.gerrard@canterbury.ac.nz (J.A. Gerrard).
with tofu quality (Poysa, Woodrow, & Yu, 2006); thus,
we were optimistic that manipulation of specific proteins
within the tofu matrix may lead to new methods to
improve product quality. By adding a cross-linking
reagent, in this case an enzyme, transglutaminase (and, in
the accompanying paper (Yasir, Sutton, Newberrry,
Andrews, & Gerrard, in press), Maillard cross-linking
agents, glutaraldehyde, formaldehyde and glyceraldehyde),
before and after the soy proteins are denatured during tofu
manufacture, it was anticipated that the textural properties
of the tofu would change to various degrees. Thus, the rela-
tionship between the degree of cross-linking of particular
sub-units and tofu texture could be established.
The consumption of soy foods in the human diet has
increased because of their beneficial effects on nutrition
and health (Friedman & Brandon, 2001). This is due to
the presence of a near perfect balance of all the essential
amino acids, making soy a valuable protein source.
Recently, the US Food and Drug Administration (FDA)
approved a health claim that soy protein reduces the risk
of coronary heart disease (Stewart, 2005). From a product

0308-8146/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.foodchem.2007.02.026
1492 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501
development point of view, soy protein has been used in a
wide range of food applications, mainly due to its func-
tional properties. Products, such as texturised vegetable
protein, soy cheese, soy yogurt and meat analogues, are
available as alternative foods. All of these products are
manufactured by manipulating soy proteins under various
conditions to meet specific functional properties of foods.
In order to achieve this, numerous studies have been under-
taken with a view to manipulating the physicochemical
properties of soy proteins (Stewart, 2005), including the
use of transglutaminase as a cross-linking agent.
Soy proteins are known to be good substrates for trans-
glutaminase (Babiker, 2000; Ikura, Kometani, Sasaki, &
Chiba, 1980; Tang, Wu, Chen, & Yang, 2006). Kwan and
Easa (2003) demonstrated that transglutaminase improved
the quality of retort tofu. Addition of transglutaminase was
found to suppress retort-induced water release, producing
a harder product. In this research, we sought to extend
understanding of the mechanism by which transglutamin-
ase improves tofu texture by extracting proteins from trea-
ted tofu samples and comparing the cross-linking that
takes place within the food matrix to that which takes place
in in vitro studies. We aimed to correlate the cross-linking
of specific soy proteins with specific attributes of the gelled
product. In this way, we hope to assist tofu manufacturers
in designing specialist tofu with particular textural proper-
ties, perhaps by varying the point of addition of the trans-
glutaminase to the processing regime.
We describe, herein, the macroscopic and microscopic
changes that take place on addition of transglutaminase
to tofu at different points during tofu manufacture and
seek to correlate these with specific cross-linking patterns.
2. Materials and methods
2.1. Materials
Unless otherwise stated, all chemicals, reagents and sol-
vents were obtained from Sigma–Aldrich New Zealand
Ltd. (Auckland, New Zealand) or BDH Chemicals New
Zealand, Ltd. (Palmerston North, New Zealand) and were
of analytical grade. Transglutaminase was obtained from
Amcor Trading Pty. Ltd. (Sydney, Australia). The com-
mercial preparation consisted of 20% protein and 80% dex-
trin filler. Defatted soy flour (52% protein content), bovine
serum albumin and bromophenol blue were purchased
from Sigma Chemical Company Ltd. (MO, USA). Soy-
beans of Chinese origin were purchased in a single batch
from the Asian Food Warehouse, Christchurch, New Zea-
land and were stored at 20 °C. Anti-foaming agent (BDH
1510 Silicone Antifoam) was purchased from BDH Labo-
ratory Supplies (Poole, England).
2.2. Protein extraction from defatted soy flour
Proteins were fractionated into glycinin and b-conglyci-
nin fractions, according to the method of Peterson and
Wolf (1992). Extractions were routinely carried out on a
10 g scale. All extractions were carried out in duplicate.
2.3. In vitro incubations
Defatted soy flour (6 mg in 600 ll of distilled water), gly-
cinin and b-conglycinin (2 mg in 600 ll of distilled water)
solutions were prepared in Eppendorf tubes. To these, an
appropriate volume of transglutaminase from a 5000 ppm
stock solution was added to obtain final concentrations
of 500 ppm and 1000 ppm. The final volume was 1 ml.
Incubation was carried out at 20 °C. Samples were
removed at required intervals. A 20 ll aliquot was pipetted
out at each interval and immediately cooled in ice water.
All treatments were carried out in duplicate.
2.4. Standard tofu preparation
The standard procedure of tofu manufacture was based
on the method of Cai, Chang, Shih, Hou, and Ji (1997).
Dried soybeans (150 g) were soaked with water for 10 h
at room temperature. After soaking, the wet beans were
rinsed. The wet beans were weighed (typically 326 g) and
subsequently blended with 730 ml of warm water (25–
30 °C) using a kitchen blender (Kambrook) at high speed
for 2 min. After blending, the slurry was juiced with a juice
extractor (Braun) and then filtered through a muslin cloth
to obtain the soymilk. The volume of soymilk was mea-
sured and found to be in the range 740–750 ml. To the soy-
milk, 20–30 ml of water were added to make the volume up
to 770 ml and one drop of anti-foaming agent was also
added. Subsequently, this soymilk was heated in a beaker,
with constant stirring, to 97 °C, on a Chiltern magnetic
stirrer hotplate (Chiltern, Auckland, New Zealand) with
the heat set at 7 °C. During heating, the beaker was cov-
ered with aluminium foil to minimise evaporation. When
the temperature of the soymilk reached 97 °C, the temper-
ature was held for 5 min before the beaker was removed
from the hot plate, and allowed to cool to 87 °C. Upon
reaching 87 °C, 750 ml of hot soymilk were measured (if
the sample was less than 750 ml, hot water was added to
make the volume up to 750 ml) and poured in to a beaker
simultaneously with 50 ml of coagulant (containing 6.7 g
CaSO4  2H2O in 50 ml of water). The coagulation was left
for 8 min to allow the soymilk to form a gel. This gel was
then transferred into the mould (Section 2.4) and com-
pressed with a 600 g (or 900 g) load for 1 h. After removing
the load, the tofu was allowed to remain standing in the
mould and left overnight at room temperature, prior to fur-
ther testing. All experiments were carried out in duplicate.
2.5. Mould design
The prototype tofu mould design was adapted from
Byun, Kang, and Mori (1995) but modified to include a
lower drain hole, which improved the consistency of the
product along the height of the column. The volume,
 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501
weight, water and protein contents in soymilk, tofu and
whey were monitored and it was verified that the mass bal-
ance was preserved. The moisture content of the sample
was determined by a gravimetric method (AOAC, 1984).
Protein contents of samples were determined by the Dumas
Method, using a LECO CNS-2000 Analyser (Laboratory
Equipment Corporation Ltd., MI, USA) calibrated with
EDTA for % nitrogen. Instrument output was in % nitro-
gen and was converted to % protein by using conversion
factor of 6.25 (AOAC, 1996). Protein content was calcu-
lated on a dry weight basis. The density of tofu was deter-
mined by measuring water displacement using a measuring
cylinder.
2.6. Preparation of transglutaminase-treated tofu
The standard tofu recipe required modification to test
the impact of transglutaminase, since the enzyme is not
active above 70 °C (Motoki & Seguro, 1998). Control tofu
for transglutaminase treatments (tga-control tofu) was pre-
pared as for standard tofu (Section 2.3) with the following
modifications: (i) 770 ml of soymilk were heated with con-
stant stirring to 50 °C, held at this temperature for 8 min
before resuming heating to 97 °C, with constant stirring,
and being held at this temperature for 5 min; (ii) after heat-
ing, the hot soymilk was allowed to cool to 50 °C. At this
temperature, 750 ml of soymilk were measured (if neces-
sary, the volume was made up with water) and poured
simultaneously into a beaker with 50 ml of coagulant.
For transglutaminase-treated tofu, the enzyme was
added, either before or after the soy milk was boiled. When
added before soymilk boiling, 32 mg of transglutaminase
(1000 ppm) or 160 mg transglutaminase (5000 ppm)
(weight transglutaminase per weight soy proteins) were
added to 770 ml of soymilk at 50 °C and held for 8 min.
Subsequently, the heating, coagulating, compressing and
storing were as for the tga-control tofu. For the transgluta-
minase added after soymilk boiling, the enzyme was added
when the soymilk had cooled to 50 °C, together with 50 ml
of coagulant and held for 8 min. The control samples had
identical thermal histories to the treated samples in each
case.
2.7. Assessment of product quality
Changes in tofu texture were used to determine the affect
of TGA on tofu product quality. Tofu texture was mea-
sured, using an Instron Universal Testing Machine (Model
4444, Canton, MA, USA) interfaced to a PC. A pro-
gramme written in LabVIEW Version 7.11 (National
Instruments, Austin, TX, USA), running on the interfaced
PC, was used to control the UTM and to collect the force
and time data during tofu compression. A flat circular
probe of diameter 65 mm, attached to the UTM crosshead,
fitted with a 50 N load cell, was used to compress a tofu
sample 36 mm in diameter and 30 mm thick. The probe
was driven down at 100 mm/min until the tofu sample
1493
had been compressed by 50% (corresponding to 15 mm
compression of tofu sample). The force–distance data were
analysed in MS Excel, using a specifically developed Visual
Basic for Applications (Microsoft, Seattle, WA, USA)
macro to determine the maximum force, compression dis-
tance and maximum compression modulus (Sharma, Mul-
vaney, & Rizvi, 2000). The compression probe and UTM
base plate were lubricated with glycerine. Three tofu sam-
ples were cut from each of the tofu columns obtained from
the mould (see Section 2.5) by cutting the column into top,
middle and bottom portions. A sample was cut from the
centre of each portion, using a 36 mm diameter circular
cutter, and then compressed on the UTM.
2.8. Protein extraction from tofu
Approximately 1 g of soybean powder, soymilk or tofu
sample was placed in an extraction thimble (30 mm 
100 mm (Whatman, Maidstone, UK)) and defatted using
100 ml of n-hexane in a standard laboratory Soxhlet appa-
ratus for about 4 h (Lusas & Riaz, 1995). The remaining
samples were dried overnight under aeration in a fume cup-
board prior to electrophoresis. Four milligrams of each
sample (defatted soybean, tofu and soymilk, and whey)
were dissolved in a 500 ll urea solution, containing 8 M
urea, 0.05 M Tris, 0.05 M dithiothrietol (DTT) at pH 10,
to extract the proteins (Woods & Orwin, 1987). The solu-
tion was then vortexed (Heidolph, Reax top, Germany),
sonicated (Sonics, Vibra cell, USA) for 15 s and centrifuged
at 11,000 rpm at room temperature for 5 min.
2.9. Analysis of soy proteins
Protein concentrations were assessed using the Bradford
(1976) method. Samples were analysed by SDS–PAGE, as
described previously for wheat proteins (Gerrard et al.,
2001). Three milligrams of defatted soy flour or 1 mg of
fractionated glycinin and b-conglycinin, with or without
transglutaminase treatment, were dissolved in 500 ll of dis-
tilled water. The solutions were vortexed and centrifuged at
14,000 rpm for 5 min at 20 °C. All SDS–PAGE analyses
were carried out in triplicate. SDS–PAGE was carried
out under reducing conditions for each extract, using pre-
cast Tris-glycine i-gels (Gradipore, NSW) at 4–20% acryl-
amide concentration. The gel image was analysed by Pho-
retix 1D Plus Software (NonLinear Dynamics Ltd.,
Newcastle-upon-Tyne, UK) for quantification of protein
subunits. The areas of protein subunits were marked and
the interference backgrounds were subtracted from the
image. All measurements were carried out in triplicate.
2.10. Scanning electron microscopy (SEM)
A Leica S440 electron microscope (Wetzlar, Germany)
was used for microscopic scanning of tofu samples, follow-
ing Kang et al. (1994). A freeze-fracture technique was
employed to observe the internal microstructure (Goldstein
1494 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501
et al., 1992). A small piece of tofu sample (<2 mm cube)
was cut with a razor blade, and immediately immersed in
liquid nitrogen for freeze-fracture. This fractured sample
was subsequently freeze-dried overnight. After freeze-dry-
ing, the sample was mounted on an aluminium stub with
double sided carbon tabs. The surface of the fractured tofu
was positioned facing up. Then, the sample was earthed
with conductive carbon paint. The sample was put in the
Polaron, sputter-coated at 1.2 kV and 20 mA for 2 min,
placed under the microscope at 10 kV and 50 pA, and at
20 mm working distance. The examination was carried
out in five replications for each treatment.
2.11. Statistical analysis
The significance of differences between means was deter-
mined by paired sample t-test using the SPSS statistical
package (Minitab Inc.). The level of significance used was
5%.
3. Results and discussion
3.1. Transglutaminase-catalysed cross-linking of soy proteins
in vitro
3.1.1. General
Before transglutaminase was introduced into the tofu
system, its effects on extracted soy protein fractions incu-
bated in vitro were examined. Transglutaminase was incu-
bated with samples of defatted soy flour and isolated
glycinin and b-conglycinin fractions, extracted from soy
MW
(kDa)
205
116
97
84
66
55
45
M Ci 0 0.25
Incubation time (h)
Fig. 1. A typical SDS–PAGE profile of defatted soy flour incubated with 1000 ppm TGA; M, marker; Ci, control initial; Cf, control final; incubation
conditions: 6 mg of defatted soy flour were dissolved in 1 ml of water and incubated with 1000 ppm of TGA at 20 °C. The image is a representative of
duplicate gels.
flour according to the method of Peterson and Wolf
(1992). Preliminary trials were undertaken to establish
standard conditions; these confirmed that soy proteins
are good substrates for transglutaminase (Ikura et al.,
1980; Tang et al., 2006).
3.1.2. Defatted soy flour
The SDS–PAGE profile of defatted soy flour incubated
with 1000 ppm TGA is presented in Fig. 1. Cross-links
between soy proteins were seen to form and the intensity
of the a0 , a and b subunits of b-conglycinin dropped rap-
idly at 0 h incubation. The intensities of A subunits were
also reduced, followed by the B subunits, which were least
reactive.
Within the B subunits, the B3 polypeptide was cross-
linked at 1000 ppm transglutaminase, but the B1, B2 and
B4 polypeptides remained throughout the incubation. This
was in agreement with the results reported by Tang et al.
(2006). The changes of intensities were quantified densito-
metrically. Reaction was sufficiently rapid that, a time zero
measurement was hard to make, with 10% and 27% of pro-
teins cross-linked before a measurement could be taken.
The concentration of unreacted protein, for all treatments,
was analysed in order to provide a point of comparison for
proteins extracted from transglutaminase-treated tofu. The
protein concentration followed an exponential decline.
The differences in reactivities of the subunits relate to the
amino acid sequence specificity of transglutaminase
(Grootjans, Groenen, & deJong, 1995; Kamiya, Takazawa,
Tanaka, Ueda, & Nagamune, 2003) and the native fold of
the proteins (Tang et al., 2006).
Aggregates
α′
α
β
A3
A1,2,4,5
B3
B1, B2, B4
A6
0.5 0.75 1.00 1.25 1.50 Cf
 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501
MW
(kDa)
205
116
97
84
66
55
45
M Ci 0 0.25 0.50
Incubation time (h)
Fig. 2. A typical SDS–PAGE profile of glycinin incubated with 1000 ppm TGA; M, marker; Ci, control initial; Cf, control final; incubation condition:
2 mg glycinin were dissolved in 1 ml of water and incubated with 1000 ppm of TGA at 20 °C. The image is a representative of duplicate gels.
3.1.3. Glycinin
Parallel analyses were carried out on an isolated fraction
of glycinin protein. After 15 min of incubation, a rapid
cross-linking of A and B subunits was observed. In the
1000 ppm incubation (Fig. 2), A and B subunits were com-
pletely cross-linked after 30 min of incubation. This sug-
gests that denaturation of the proteins, during extraction
from the defatted soy flour (enhanced by the 2-mercap-
toethanol present in the extraction buffer), made the B
sub-units more accessible to transglutaminase-catalysed
cross-linking than had been the case in defatted soy flour.
The formation of cross-linked proteins was not observed
on the top of the gel, due to the formation of very large
aggregates that were unable to penetrate the matrix.
3.1.4. b-Conglycinin
Analogous experiments, using the b-conglycinin fraction
extracted from defatted soy flour incubated with 500 and
1000 ppm TGA, showed similar results. At 1000 ppm
TGA, especially, the intensities of a0 , a and b subunits were
rapidly reduced (Fig. 3). After 30 min of incubation, these
subunits were mostly cross-linked, in agreement with the
results reported by Tang et al. (2006). Since the b-conglyc-
inin fraction was cross-contaminated with A and B
subunits of glycinin, the A subunits were also rapidly
cross-linked. Within the B subunits, the B3 polypeptide
rapidly cross-linked, but the B1, B2 and B4 polypeptides
remained unreacted. This profile is similar to that observed
in defatted soy flour (Fig. 1). The gels were analysed densi-
tometrically in order to provide a point of comparison for
the transglutaminase-treated tofu.
The change in reactivity with transglutaminase upon
denaturation has implications for the reactivity of transglu-
1495
Aggregates
α′
α
β
A3
A1,2,4,5
B1,2,3,4
A6
0.75 1.00 1.25 1.50 Cf
taminase within the tofu matrix, where the denaturation of
the native proteins will vary according to the point in pro-
cessing at which the enzyme is added.
3.2. Comparison of fracture force and microstructure for
standard tofu to tga-control tofu
3.2.1. Fracture force of control tofus
The standard tofu manufacturing process (Cai et al.,
1997) gave a standard tofu with properties comparable to
commercial Chinese tofu, purchased from a local super-
market. In order to assess the impact of transglutaminase
on the texture of tofu, it was necessary to modify the stan-
dard process in order to ensure that the enzyme was added
to the soy milk at a temperature cool enough to preserve its
activity. This resulted in a tga-control tofu with a signifi-
cantly lower fracture force than the standard preparation
(P < 0.05). Thus the standard tofu preparation had a frac-
ture force over 30 N, whereas the tga-control tofu had a
fracture force of less than 25 N. The differing thermal his-
tories of the standard and tga-control tofus may account
for the differing fracture behaviours of these otherwise
identical tofus.
3.2.2. Microstructure of tofu
Scanning electron microscopy (SEM) has been variously
employed to examine the microstructure of food, and relate
the structure to texture (Wilkinson, Dijksterhuis, & Mine-
kus, 2000), with uniform continuous networks showing the
firmest textures (Kao, Su, & Lee, 2003). Freeze-fracture
SEM was used to capture high magnification images of
the internal microstructure of standard and tga-control
tofu.
1496 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501
MW
(kDa)
205
116
97
84
66
55
45
M Ci 0 0.25
Incubation time (h)
Fig. 3. A typical SDS–PAGE profile of b-conglycinin incubated with 1000 ppm TGA at 20 °C; M, marker; Ci, control initial; Cf, control final; a0 , a and b
are subunits of b-conglycinin. A and B are the cross-contaminated acidic and basic subunits of glycinin. The image is a representative of duplicate gels.
The microstructure of standard tofu (Fig. 4A) showed
small pores distributed uniformly in a dense, well con-
nected network. In contrast, the tga-control tofu
(Fig. 4B) had larger pores, and consisted of a ‘‘flakey”, dis-
continuous, loose network, with wide pores and a diffuse
structure. The appearance of the tofu gel microstructure
was consistent with the observed changes in fracture force.
3.3. Analysis of microstructure and fracture force for
transglutaminase-treated tofu
Preliminary experiments suggested that 1000–5000 ppm
of transglutaminase were required in order to change the
firmness of tofu. Samples of tofu were prepared, with trans-
glutaminase added either before the soy milk was boiled, or
after the soy milk had been boiled and cooled to the appro-
priate temperature. The microstructures of tofu treated
with 1000 ppm and 5000 ppm transglutaminase, before
and after boiling, are shown in Fig. 4. The corresponding
fracture forces are illustrated in Fig. 5.
Compared to the control, the fracture force at both
transglutaminase concentrations, whether added before or
after soymilk boiling, increased. Further, with the increase
in transglutaminase concentration, both for treatment
before and after boiling, the fracture force significantly
increased (P < 0.05). However, the fracture force for the
tofu treated after boiling was lower than that for those
treated before boiling, perhaps reflecting the limited access
of the enzyme to its substrates, which had already begun to
form a gel. To corroborate the fracture force measure-
ments, the compression modulus of the samples was also
examined. The compression modulus followed the same
pattern as the fracture force: for all transglutaminase con-
Aggregates
α’
α
β
A3
A1,2,4,5
B3
B1,B2,B4
A6
0.50 0.75 1.00 1.25 1.50 Cf
centrations, it was significantly increased (P < 0.05) over
the tga-control sample. Moisture content was measured
for all samples, but no correlation was found with the firm-
ness of tofu.
Using 1000 ppm of transglutaminase before boiling
(Fig. 4C), the microstructure showed a fine-stranded net-
work with small, uniformly distributed pores. However,
when the concentration of transglutaminase was increased
to 5000 ppm (Fig. 4D), the microstructure did not show
this ‘honey comb’ type structure, and the aggregates
seemed more rigid. With transformation from a loose net-
work, in the tga-control tofu, to a fine-stranded network,
to a more rigid structure, the fracture force of tofu
increased. The fracture force of 1000 ppm transglutamin-
ase-treated tofu was similar to the standard tofu, reflecting
their similar microstructures.
A similar trend in microstructure was observed with
treatment after boiling. With 1000 ppm transglutaminase
(Fig. 4E), the microstructure showed a ‘‘sponge-like” struc-
ture. This resulted in the lowest fracture force among all
transglutaminase treatments. When the concentration of
TGA was increased to 5000 ppm (Fig. 4F), a fine-stranded
network was also formed, but was ‘tighter’ in appearance.
This structure corresponded to a tofu with a greater frac-
ture force, although not significantly different from the
sample treated with the same concentration of enzyme
before boiling.
Thus transglutaminase treatment promoted the forma-
tion of fine-stranded networks and gave a more homoge-
neous structure with a higher fracture strength. In all
cases, a firmer tofu was obtained, especially for the samples
treated before boiling. This suggests that transglutaminase
treatment may offer a means to increase the firmness of
 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501 1497

Fig. 4. Freeze-fracture SEM micrographs of standard and TGA-treated tofu. Both control tofus, the standard control tofu (A), and the TGA-control (B)
are shown. Tofu was treated with TGA added either before boiling (at 1000 ppm) (C) and 5000 ppm (D) or after boiling (at 1000 ppm (E) and 5000 ppm
(F)). The images are representative of five observations of replicate experiments. The black scale bar represents 10 lm.

40
tofu, to a degree that can be influenced by enzyme dose and
addition point. 35

30
3.4. SDS–PAGE analysis of protein extracts from
Fracture Force (N)

transglutaminase-treated tofu 25

Since transglutaminase had been shown to cross-link soy 20

proteins in vitro and to alter the texture of tofu in situ, we 15
expected to find that cross-linking had occurred, within the
tofu matrix, to varying degrees, according to the degree of 10
denaturation of the soy proteins. Proteins were thus
5
extracted from control and transglutaminase-treated tofu
in order to establish whether specific cross-linking patterns 0
might account for the observed changes in fracture force 1 2 3 4 5
and microstructure. To our surprise, when the extracted
Fig. 5. Fracture forces of transglutaminase-treated tofu: 1 – 1000 ppm
proteins were analysed by SDS–PAGE, only very limited transglutaminase added before boiling; 2 – 1000 ppm of transglutaminase
cross-linking occurred in tofu that had been treated with added before boiling; 3 – tga-control; 4 – 1000 ppm of transglutaminase added
transglutaminase, either before or after boiling. after boiling; and 5 – 5000 ppm of transglutaminase added after boiling.
1498 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501

Soymilk Tofu Whey

MW
Aggregates
(kDa)

205
α′
α

β
116
A3
97 A1,2,4,5

84

66 B1,2,3,4

55 A6

M 1 2 3 M 1 2 3 M 1 2 3

Fig. 6. SDS–PAGE profile of proteins in soymilk, tofu and whey under reducing conditions in the treatment before boiling. M = marker; 1 – tga-control;
2 and 3 are 1000 and 5000 ppm of TGA. The electrophoretic profiles are representative of duplicate gels.

Fig. 6 shows the SDS–PAGE analysis of proteins tometry revealed that the relative concentration of proteins
extracted from soy milk and tofu samples that had been decreased to 77% of the control value at 1000 ppm, and
treated with transglutaminase before boiling, and Table 1 52% at 5000 ppm. However, the total concentration of
shows the densitometric analysis of these gels. No signifi- these proteins was very low, and did not significantly
cant loss of any sub-unit was apparent at either 1000 or impact on the protein content of the treated tofu.
5000 ppm transglutaminase treatment levels. A decrease The SDS–PAGE analysis of the proteins of soymilk,
in aggregated material was recorded for the tofu treated tofu and whey, for samples treated with transglutaminase
with 5000 ppm of transglutaminase, perhaps reflecting the after boiling, are presented in Fig. 7 and Table 2. Since
further aggregation of already aggregated material, to a the soymilk sample had not been treated with transgluta-
size that could no longer enter the gel. minase, no change in protein profile was expected. In the
In the corresponding whey samples, lower molecular transglutaminase-treated tofu samples, no loss of protein
weight protein bands were observed in a range of 14– subunits was seen at 1000 ppm, but, at 5000 ppm TGA,
24 kDa. As the concentration increased from 1000 to the intensities of the subunit bands decreased slightly, to
5000 ppm, a number of bands started to disappear, sug- 68–98% of the control, suggesting that small proportions
gesting that lower molecular weight proteins were the only of proteins were cross-linked. These results suggest that
proteins to be cross-linked by the enzyme in tofu. Densi- the cross-linking activity of the enzyme was increased,
when added after boiling the soymilk, than when added
before boiling, consistent with greater substrate availability
in the denatured protein. The cross-linking reaction
Table 1 appeared to favour the smaller subunits, again hinting at
The changes of protein concentration in the cross-linked proteins and a preferential availability of smaller substrates to the
subunits in the transglutaminase treatment, with the enzyme added before
enzyme in the protein mixture.
soymilk boiling
In the corresponding whey, the electrophoretic profile
Aggregates and Relative protein concentration (% of tga-control)
was in contrast to those obtained for treatment before boil-
subunits
Soymilk Tofu ing. As the transglutaminase concentration was increased,
1000 ppm 5000 ppm 1000 ppm 5000 ppm the number of low molecular weight proteins present in
Aggregates 100 ± 2 96 ± 1 95 ± 3 57 ± 4 the whey increased to 130% of the control for the
a0 + a 95 ± 9 85 ± 16 95 ± 5 97 ± 1 5000 ppm transglutaminase treatment. This perhaps indi-
b 92 ± 8 84 ± 14 92 ± 9 95 ± 15 cates that transglutaminase had increased the solubility
A3 86 ± 12 72 ± 21 93 ± 2 83 ± 3
of the whey proteins, as has been previously noted for
A1,2,4,5 91 ± 5 79 ± 13 95 ± 4 100 ± 1
B 88 ± 1 79 ± 16 95 ± 4 98 ± 11 soy proteins (Babiker, 2000) and postulated to be due to
A6 89 ± 5 69 ± 26 81 ± 4 93 ± 18 the known side activity of transglutaminase, hydrolysis of
Values are the means ± standard error of the means of duplicate mea- glutamine residues to glutamic acid (Motoki, Seguro,
surements. The percentage is based on the control soymilk and tga-control Nio, & Takinami, 1986). Why this should be different to
tofu. the before boiling treatment is not clear.
 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501 1499

Soymilk Tofu Whey

MW
Aggregates
(kDa)

205
α′
α

β
116
A3
97 A1,2,4,5
84

66 B1,2,3,4

55
A6

Fig. 7. SDS–PAGE profile of proteins in soymilk, tofu and whey under reducing conditions in the treatment after boiling. M = marker; 1 = control; 2 and
3 are 1000 and 5000 ppm of TGA; soymilk was not treated with TGA. The electrophoretic profiles are representative of duplicate gels.

Table 2 4. Conclusion
The changes of protein concentration in the cross-linked proteins and
subunits in the transglutaminase treatment, with the enzyme added after Treatment with transglutaminase produced a firmer
soymilk boiling tofu, with a significantly increased fracture force, suggest-
Cross-linked and subunits Relative protein concentration in tofu (% of ing that it may indeed be a useful ingredient for tofu man-
tga-control) ufacturers. SEM revealed that the microstructures of the
1000 ppm TGA 5000 ppm TGA samples were consistent with the changes in fracture force.
Aggregates 91 ± 6 80 ± 7 The firmness varied according to both the concentration of
a0 + a 92 ± 3 74 ± 5 enzyme added and the point of addition in the manufactur-
b 100 ± 1 88 ± 9 ing process, suggesting opportunities for customising tofu
A3 86 ± 11 68 ± 17
A1,2,4,5 100 ± 1 79 ± 8
by simple alterations in the timing of transglutaminase
B1,2,3,4 100 ± 1 90 ± 5 addition.
A6 93 ± 8 85 ± 11 At the molecular level, in vitro cross-linking of soy pro-
Values are the means ± standard error of the means of duplicate mea- teins resulted in efficient cross-linking of most proteins in
surements. The percentage is based on the control tofu, respectively. the extract, confirming earlier research that showed soy
proteins to be an excellent substrate for transglutaminase
Taken as a whole, these results suggest that, whilst (Tang et al., 2006). However, when proteins were extracted
transglutaminase clearly alters the microstructure and tex- from the tofu matrix after treatment with transglutaminase
ture of tofu, cross-linking does not play a major role in the in situ, little, if any, cross-linking occurred. This observa-
mechanism of action. tion challenges the tacit assumption made by many

(i)

N TGA O-
+ +
H3 N
O H2O O

ε-(γ-glutamyl)lysyl cross-link Glutamate Lysine

(ii)
NH 2
TGA O-
+ NH3
O H2O = Protein
O

Glutamyl residue Glutamate Ammonia

Fig. 8. The hydrolysis of glutamyl residue of the protein by TGA, either by hydrolysis of glutamyl-lysyl cross-links (i), or glutamyl residues (ii) (Motoki
et al., 1986).
1500 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501
researchers that the results from in vitro incubations of
food components with food enzymes can be assumed to
be valid within the food.
These findings suggest that the change in functional
properties afforded by addition of transglutaminase to tofu
is perhaps due to a side reaction of the enzyme, hydrolysis
of glutamine residues to glutamate, as previously postu-
lated by Babiker (2000). Since water is abundant during
tofu manufacture, it is likely that transglutaminase-pro-
moted hydrolysis of glutamine residues of the protein to
glutamate residues could occur, by one of two mechanisms,
as shown in Fig. 8. Either the e-(c-glutamyl)lysyl cross-link
is formed and subsequently hydrolysed into glutamate and
lysine, or the glutamyl residue is hydrolysed directly into
glutamate. The resulting change in isoelectric point of the
soy proteins would be predicted to change the gelation
and microstructure of the proteins (Dickinson, 1997). Fur-
ther work is needed to test this hypothesis in this, and
other, systems in which side reactions of transglutaminase
are implicated as important for changing the properties
of food (Gerrard & Sutton, 2005).
In order to further probe the relationship between cross-
linking and tofu texture, an alternate means of cross-link-
ing of the protein gel matrix was explored: the introduction
of Maillard cross-linking reagents. This work is described
in the following paper (Yasir et al., in press).
Acknowledgements
This work was funded by the Malaysian Government, as
a scholarship to S.B.Y. We thank Jackie Healy for tena-
cious technical assistance, and Drs. Sean Devenish and
Grant Pearce for very critical comment on an earlier ver-
sion of the manuscript.
References
AOAC (1984). Official method of analysis. Washington, DC: Association
of Official Analytical Chemists.
AOAC (1996). Official methods of analyses. Washington, DC: Association
of Official Analytical Chemists.
Babiker, E. E. (2000). Effect of transglutaminase treatment on the
functional properties of native and chymotrypsin-digested soy protein.
Food Chemistry, 70, 139–145.
Bradford, M. M. (1976). A rapid and sensitive method for the
quantification of microgram quantities of protein utilising the
principle of protein dye binding. Analytical Biochemistry, 72, 248–
254.
Byun, M. W., Kang, I. J., & Mori, T. (1995). Properties of soya milk and
tofu prepared with c-irradiated soya beans. Journal of the Science of
Food and Agriculture, 67, 477–483.
Cai, T. D., Chang, K. C., Shih, M. C., Hou, H. J., & Ji, M. (1997).
Comparison of bench and production scale methods for making
soymilk and tofu from 13 soybean varieties. Food Research Interna-
tional, 30, 659–668.
Dickinson, E. (1997). Enzymatic cross-linking as a tool for food colloid
rheology control and interfacial stabilization. Trends in Food Science
and Technology, 8, 334–339.
Friedman, M., & Brandon, D. L. (2001). Nutritional and health benefits of
soy proteins. Journal of Agricultural and Food Chemistry, 49,
1069–1086.
Gerrard, J. A. (2002). Protein–protein cross-linking: Methods, conse-
quences, applications. Trends in Food Science and Technology, 13,
389–3973.
Gerrard, J. A., Fayle, S. E., Brown, P. K., Sutton, K. H., Simmons, L., &
Rasiah, I. (2001). Effects of microbial transglutaminase on the wheat
proteins of bread and croissant dough. Journal of Food Science, 66,
782–786.
Gerrard, J. A., Fayle, S. E., Wilson, A. J., Newberry, M. P., Ross, M., &
Kavale, S. (1998). Dough properties and crumb strength of white pan
bread as affected by microbial transglutaminase. Journal of Food
Science, 63, 472–475.
Gerrard, J. A., Newberry, M. P., Ross, M., Wilson, A. J., Fayle, S. E., &
Kavale, S. (2000). Pastry lift and croissant volume as affected by
microbial transglutaminase. Journal of Food Science, 65, 312–
314.
Gerrard, J. A., & Sutton, K. H. (2005). Addition of transglutaminase to
cereal products may generate the epitope responsible for
coeliac disease. Trends in Food Science and Technology, 16, 510–
512.
Grootjans, J. J., Groenen, P. J. T. A., & deJong, W. W. (1995). Substrate
requirements for transglutaminases – Influence of the amino acid
residue preceding the amine donor lysine in a native protein. Journal of
Biological Chemistry, 270, 22855–22858.
Goldstein, J. I., Newbury, D. E., Echlin, P., Joy, D. C., Romig, A. D. J.,
Lyman, C. E., et al. (1992). Scanning electron microscopy and X-ray
microanalysis. New York: Plenum Press.
Ikura, K., Kometani, T., Sasaki, R., & Chiba, H. (1980). Cross-linking of
soybean 7S and 11S proteins by transglutaminase. Agricultural and
Biological Chemistry, 44, 2979–2984.
Kamiya, N., Takazawa, T., Tanaka, T., Ueda, H., & Nagamune, T.
(2003). Site-specific cross-linking of functional proteins by transgluta-
minase. Enzyme and Microbial Technology, 33, 492–496.
Kang, I. J., Matsumura, Y., Ikura, K., Motoki, M., Sakamoto, H., &
Mori, T. (1994). Gelation and gel properties of soybean glycinin in a
transglutaminse-catalysed mixture. Journal of Agricultural and Food
Chemistry, 42, 159–165.
Kao, F. J., Su, N. W., & Lee, M. H. (2003). Effect of calcium sulfate
concentration in soymilk on the microstructure of firm tofu and the
protein constitutions in tofu whey. Journal of Agricultural and Food
Chemistry, 51, 6211–6216.
Kwan, S. W., & Easa, A. M. (2003). Comparing physical properties of
retort-resistant glucono-d-lactone tofu treated with commercial trans-
glutaminase enzyme or low levels of glucose. Lebensmittel-Wissens-
chaft & Technologie, 36, 643–646.
Liu, K. S. (1997). Soybeans chemistry: Technology and utilization. New
York: Chapman & Hall.
Lusas, E. W., & Riaz, M. N. (1995). Soy protein products: Processing and
use. Journal of Nutrition, 125, 573S–580S.
Motoki, M., & Seguro, K. (1998). Transglutaminase and its use
for food processing. Trends in Food Science Technology, 9,
204–210.
Motoki, M., Seguro, K., Nio, N., & Takinami, K. (1986). Glutamine-
specific deamidation of j-casein by transglutaminase. Agricultural and
Biological Chemistry, 50, 3025–3030.
Peterson, R. E., & Wolf, W. J. (1992). Enhancement of high-performance
liquid chromatograph of soybean proteins by addition of sodium
dodecyl sulfate. Cereal Chemistry, 69, 101–102.
Poysa, V., Woodrow, L., & Yu, K. (2006). Effect of soy protein subunit
composition on tofu quality. Food Research International, 39,
309–317.
Sharma, S. K., Mulvaney, S. J., & Rizvi, S. S. H. (2000). Food
processing engineering: Theory and laboratory experiments. New
York: Wiley.
Stewart, K. L. (2005). Soy good for you. Better Nutrition, 67, 24–26.
Tang, C.-H., Wu, H., Chen, Z., & Yang, X.-Q. (2006). Formation and
properties of glycinin-rich and b-conglycinin-rich soy protein isolate
gels induced by microbial transglutaminase. Food Research Interna-
tional, 39, 87–97.
 S. Bin Md Yasir et al. / Food Chemistry 104 (2007) 1491–1501 1501

Wilkinson, C., Dijksterhuis, G. B., & Minekus, M. (2000). From food Yasir, S. Bin Md, Sutton, K. H., Newberrry, M. P., Andrews, N. R., &
structure to texture. Trends in Food Science and Technology, 11, 442–450. Gerrard, J. A. (in press). The impact of Maillard cross-linking on soy
Woods, J. L., & Orwin, D. F. G. (1987). Wool proteins of New Zealand proteins and tofu texture. Food Chemistry, doi:10.1016/
Romney sheep. Australian Journal of Biological Science, 40, 1–14. j.foodchem.2007.02.042.